Abstract
A growing number of geotechnical specialists use wide area InSAR to complement their toolbox of terrestrial monitoring for slope stability. The best practitioners have found ways to query the enormous InSAR data sets for the purpose of addressing geotechnical risk very early in the monitoring process. Failure is usually preceded by rapid deformation on meter scale which is preceded by slow deformation on millimeter scale which is the type of deformation assessed by InSAR.
These fuzzy data on surface deformation are noisy and when sampling stability with millimeter precision, the reality is that everything on the ground moves continuously at these velocities.
But if everything is moving, how do practitioners know which movements are of concern? In scientific publications we often see highly stylized and select areas for retro-active curve fitting. In a data set containing millions of moving points on the ground, this is a lot harder to do proactively.
The reality of noisy fuzzy data and stylized inverse velocity graphs is a hard one to match if you look at a complex site and feel the weight of geotechnical responsibility. The challenge for the geotechnical engineer in this new world is to first get this InSAR data right and then to not drown in the data. We demonstrate a multistep process in which we separate geotechnical precursor signal from other ground motion “noise”.
This presentation discusses two cases where this multi-step process of wide area InSAR data was done successfully at Oceana Gold’s Haile Mine. One case addresses a large slope with structural issues, where the GB-SAR field of view was insufficient to tell if the slope as a whole was standing still … or was moving. Another was a slope instability developing after severe rainfall: would this part of the mine need to be closed to operations or could it remain open